Universidad Carlos III de Madrid

Econometrics
ME-MIEM
Instrumental Variables Regression - II
Problem Set 9

1. In an instrumental variables regression model with a single regressor Xi ; and a single instrument Zi ; the
regression of Xi on Zi has R2 = 0:05 and n = 100: Is Zi an strong instrument? Would you change your answer
if R2 = 0:05 and n = 500?

2. Use the regressions in the following table to answer the questions. For that data from 48 states of the USA
were used. The statistics were statistically signi…cant at the 5% or 1% level.
Dependent Variable: ln(Qcigarretes
i;1995 ) ln(Qcigarretes
i;1985 )
Regressor (1) (2) (3)
cigarretes cigarretes
ln(Pi;1995 ) ln(Pi;1985 ) 0:94 1:34 1:20
(0:21) (0:23) (0:20)

ln(Incomei;1995 ) ln(Incomei;1985 ) 0:53 0:43 0:05

(0:34) (0:30) (0:06)

Intercept 0:12 0:02 0:05

(0:07) (0:07) (0:06)
Sp eci…c tax Sales tax and Sp eci…c
Instrumental V ariables Sales Tax cigarettes tax cigarettes

F Statistic of the …rst stage 33.70 107.20 889.60

J T est f or restrictions
overidentif ication and p value - - 4:93
(0:026)

(a) Imagine that the federal government is considering a new tax on the cigarettes which is estimated will
increase the retail price in 0.50$ per packet. If the current price per packet is now 7.40$, use the regression
in column (1) to predict the variation in demand. Construct a 95% con…dence interval for the variation
in demand.
(b) Imagine that the USA gets in recession and the income is reduced by 2%. Use the regression in column
(1) to predict the variation in demand.
(c) Imagine that the recession last less than one year. Do you think that the regression in column (1) provides
a reliable answer to the question in (b)? Why or why not?
(d) Imagine that the F statistics in column (1) were 3.6 instead of 33.6. Would the regression provide a
reliable answer to the question set in (a)? Why or why not?

3. Consider the regression model with a single regressor

Yi = 0 + 1 Xi + Ui :

Assume that all the conditions of the classical regression model are satis…ed.

(a) Show that Xi is a valid instrument. That is, show that the conditions of relevance of the instrument and
of exogeneity of the instrument hold.
(b) Show that the assumptions of the IV regression hold with this choice of Zi :
(c) Show that the IV estimate with Zi = Xi is identical to the OLS estimate.

4. Consider the regression model with a single regressor

Yi = 0 + 1 Xi + Ui :

We are interested in estimating the variance of the error term in this equation.

1
 (a) Assume that we use the estimate of the second stage regression of 2SLS
n
X 2
1 ^ 2SLS ^ 2SLS X
^i
^ 2a = Yi 0 1 ;
n 1 i=1

^ i is the …tted value of the …rst stage. Is this estimate consistent? (For the answer to this question,
where X
assume that the sample size is very large and that the 2SLS estimates are essentially equal to 0 and 1 ):
(b) Is this estimate consistent
n
X 2
1 ^ 2SLS ^ 2SLS Xi
^ 2b = Yi 0 1 ?
n 1 i=1

5. Consider the instrumental variables regression model

Yi = 0 + 1 Xi + 2 Wi + ui ;

where Zi is an instrument. Suppose that the data on Wi are not available and that you estimate the model
omitting Wi in the regression.

(a) Suppose that Zi and Wi are not correlated. Is the IV estimate consistent?
(b) Suppose that Zi and Wi are correlated. Is the IV instrument consistent?

6. Consider the instrumental variables regression model

Yi = 0 + 1 Xi + 2 Wi + Ui

where Xi is correlated with Ui and Zi is an instrument. Assume that the following three conditions hold:

(a) E (Ui jWi ) = 0;

(b) (Xi ; Wi ; Zi ; Yi ) are i.i.d. draws of their joint distribution.
(c) Extreme large values are not very likely: the variables X; W; Z and Y have …nite moments of fourth order
and di¤erent from zero.

Which IV assumption does not hold in each of the following cases?

(a) Zi is independent of (Xi ; Wi ; Yi ) :

(b) Zi = Wi :
(c) Wi = 1 for all i:
(d) Zi = Xi :

7. The following model measures the e¤ect of a programme of school choice in terms of the score of an standardized
test,
score = 0 + 1 choice + 2 faminc + u1 ;
where score is the score of the test conducted at state level, choice is the binary variable which indicates if a
student has attended a choice school during the last year, and faminc is the family income. The IV for choice
is grant; the amount in dollars received by the students to pay the fees of the choice school. The grant received
varies over the income level of the family, which is the reason why we control for faminc in the equation.

(a) Even with faminc in the equation, why choice could be correlated with u1 ?
(b) If within each income level, the amount of the grant were randomly assigned, would grant be correlated
with u1 ?
(c) Write down the reduced form equation for choice: Which is needed for grant to be partially correlated
with choice?
(d) Write down the reduced form equation for score: Explain why it could be useful [Hint: how would you
interpret the coe¢ cient of grant?].

2
8. Suppose that the annual income and the alcohol consumption are determined by the following simultaneous
equations system,

log (earnings) = 0 + 1 alcohol + 2 educ + u1

alcohol = 0 + 1 log (earnings) + 2 educ + 3 log (price) + u2 ;

where price is a local index of alcohol prices, which include local and state taxes. Assuming that educ and
price are exogenous, and that 1 ; 2 ; 1 ; 2 and 3 are all di¤erent from zero, which equation is identi…ed?
How would you estimate this equation?

ANSWERS

1. F = 5:16, so that the instrument can be weak. With 500 obs. F = 26:2 so that the answer can change.

2. (a) -6.15% and [-8.84%, -3.46%].

(b) -1.06%.
(c) No, if recession lasts less than one year.
(d) No.

5. .

(a) Yes.
(b) No.

7. .

(a) Which factors are contained in u1 ?

(b) Yes.
(c) The coe¢ cient of grant must be di¤erent from zero.
(d) The coe¢ cient of grant would be helpful to interpret the IV estimate.

8. The …rst one.